Background Rheumatoid Arthritis (RA) is a chronic autoimmune disease with higher prevalence and worse prognosis in women than men.
Objectives We aimed to determine whether 48 single nucleotide polymorphisms (SNPs) within steroid hormone signaling (ESR1, ESR2, PGR, NR1I2, and SHBG), phase I- and II-metabolizing enzyme (HSD17B1, CYP1A1, CYP17A1, CYP1A2, CYP1B1, CYP2C9, CYP2C19, CYP3A4 and GSTP1) and hormone transporter (ABCB1) genes influence on the risk of developing RA and whether genotyping of these variants might help to predict disease risk.
Methods The study population included 1357 RA patients and 1219 controls. Logistic regression analyses were performed to determine associations and to predict disease risk.
Results We found that carriers of the CYP2C19rs4244285A allele and CYP3A4rs11773597C/C genotype had a significantly increased risk of RA (P=0.019 and P=0.0048) whereas carriers of the CYP2C9rs1057910C and ESR2rs4986938G alleles showed a reduced risk of developing the disease (P=0.036 and P=0.0026). In addition, when a log-additive model was assumed, we observed a significant association for the CYP2C19rs4244285, CYP3A4rs2740574 and ESR2rs4986938 SNPs suggesting an allele-dosage effect of these variants to modulate the risk of RA (P=0.0076, P=0.044 and P=0.0004). Interestingly, a gender-stratified analysis also revealed that women carrying the CYP17A1rs743572G/G genotype or the PGRrs518162A allele showed an increased risk of RA (P=0.026 and P=0.032) whereas an opposite but not significant effect was observed in men (Pinteraction=0.02 and 0.025, respectively). Furthermore, a rheumatoid factor (RF)-stratified analysis showed that seropositive patients carrying the SULT1A1rs9282861A allele had a substantially increased risk of developing RA (P=0.0088) whereas no effect was observed in seronegative patients (P=0.56). Importantly, after correction for multiple testing (P=0.0010), the association of the ESR2rs4986938 SNP with a reduced risk of RA remained statistically significant (P=0.0004) whereas the overall association of CYP2C19rs4244285 and CYP3A4rs2740574 or SULT1A1rs9282861 in seropositive patients reached marginal significance (P=0.0076, P=0.0048 and P=0.0088). Finally, a predictive analysis showed that a model including 3 genetic variants significantly associated with RA had a higher prediction capacity than a model including a similar number of non-significant SNPs (AUROC=0.562 vs. AUROC=0.513; -2log likehood ratio test P=4.24E-07). The predictive analysis showed that the ESR2rs4986938 SNP had the highest predictive capacity (P=0.00015), which along with a previous study demonstrating its correlation with ESR2 mRNA expression suggest a key role of this variant in the modulation of RA risk.
Conclusions These findings suggest that SNPs within estrogen-related genes may play a role in modulating susceptibility to RA and can be used to improve the prediction of disease risk.
Disclosure of Interest None declared